Application of the Algorithm for the Recognition of Pedestrian Disturbance Patterns by Lucas-Kanade Method in Real Time

The use of security cameras allows obtaining relevant information that through video processing can provide a solution to the problem that every country faces, crime. In the following work an algorithm is developed for the detection of criminal acts, taking into account the disturbance pattern, with the use of the Python 3 programming language. For the identification of people, the area occupied by the body was taken into account, which allowed to discard any other object seen in the video. On the other hand, the monitoring of people and the analysis of changes in speed through distance was the product of tracking a corner of the body identified in a set of superimposed images, this was done with the help of the OpenCV library. The results show that the proposed algorithm can be improved with the additional use of patterns to increase its precision.